Relative decay index and sources of polycyclic aromatic hydrocarbons

An optimized method for the analysis of polycyclic aromatic hydrocarbons (PAH) in atmospheric aerosols with short sampling times (1h) has been used to determine the daily variations of PAH in the atmosphere. Of the various physicochemical parameters controlling the disappearance of PAH, the most important seem to be the thermic dependence and the chemical reactions with gaseous pollutants (NO_x, O₃, HNO₃, OH). The RDI is used to identify the different sources of urban pollution: domestic heating, vehicle traffic (petrol and diesel), refineries, foundries, incinerators and power stations (coal, gas and oil) and to quantify the results for standard conditions. The characteristic concentration ratios from each source are often different from those in the literature, where the data are based on long sampling times and are affected by PAH reactivity differences. The results obtained are parameters for setting up a mathematical model for calculating concentrations of PAH at any receptor site.     

A comparison study of three urban air pollution models

The predictions of three urban air pollution models with varying degrees of mathematical and computational complexities are compared against the hourly SO₂ ground-level concentrations observed on 10 winter nights of the RAPS experiment in St. Louis. The emphasis in this study is on the prediction of urban area source concentrations. Statistics for the paired comparison of predictions of each model with the observations are presented. The RAM and the ATDL model with stable diffusion coefficients overestimated the observed night-time concentrations. The results show that the performance of the ATDL model with near-neutral diffusion coefficients is comparable to the more sophisticated 3-D grid numerical model.     

An analytical model for crosswind integrated concentrations released from a continuous source in a finite atmospheric boundary layer

An analytical model for the crosswind integrated concentrations released from a continuous source in a finite atmospheric boundary layer is formulated by considering the wind speed as a power law profile of vertical height above the ground and eddy diffusivity as an explicit function of downwind distance from the source and vertical height. A closed form analytical solution of the resulting advection–diffusion equation for these profiles of wind speed and eddy diffusivity with the physically relevant boundary conditions is derived using the separation of variables technique that leads to a Sturm–Liouville eigen value problem. Various particular cases of the model are deduced.The model is evaluated with the observations obtained from Prairie Grass experiment in various stability classes varying from very unstable to neutral and stable conditions and Hanford diffusion experiment in stable conditions. The agreement is found to be good between the computed and observed concentrations in both the diffusion experiments. For Prairie Grass experiment, the model is predicting 78% cases with in a factor of two and gives a value of NMSE as 0.075. On the other hand, for Hanford observations in stable conditions, it predicts 70% cases with in factor of two. An extensive analysis of statistical measures with the downwind distances from the source reveals that the model is performing well close to the source.     

Chemical mass balance receptor model applied to ambient C2–C9 VOC concentration in Seoul,Korea: Effect of chemical reaction losses

A chemical mass balance (CMB) receptor model was used for estimating the diurnal contributions of VOC emission sources to the ambient C₂–C₉ VOC concentration in Seoul, Korea. For this purpose, the VOC concentrations were measured in the morning, the afternoon, and the evening. The samples were collected using a 2-h integrated SUMMA canister. The source profiles were developed for the CMB calculation in the Seoul area. To investigate the effect of the chemical reaction loss of VOCs on the CMB calculation, the modified model employing a decay factor and the standard model that considers no loss were compared. The modified model estimated that the vehicle exhaust (52%) was the largest leading source of VOCs in the Seoul atmosphere, followed by the use of solvents (26%), gasoline evaporation (15%), the use of liquefied petroleum gas (LPG) (5%), and the use of liquefied natural gas (LNG) (2%). Relative source contribution for vehicle exhaust showed a clear diurnal variation with a high in the morning and evening and a low in the afternoon, while the contribution of evaporative emissions (gasoline evaporation and solvent usage) showed a different diurnal pattern from that of the vehicle exhaust, exhibiting a high in the afternoon and evening and a low in the morning. It was found that the difference between the total source contribution (μg m⁻³) estimated from these two models was not statistically significant. However, when the paired-sample t-test is applied to the individual sources, a significant difference was found for the vehicle exhaust and the solvent use. In addition, the modified model brought forth a better performance with high R² and low χ² as compared to those obtained from the standard model in the CMB calculation. The vehicle exhaust and solvent use were estimated to be the largest and the second largest contributors to ambient benzene as well as ozone formation potential (OFP), respectively. Based on above results we believe that incorporating the reaction loss in the CMB calculations helps to better fit the source profile to the ambient VOC concentrations. However, the reaction loss does not significantly affect the estimation of source contributions.     

Modeling the intra-urban variability of outdoor traffic pollution in Oslo,Norway—A GA2LEN project

Traffic is a major source of air pollutants in urban environments, and exposure to these pollutants may be associated with adverse health effects. However, inconsistencies in observational epidemiological studies may be caused by differential measurement errors in various approaches in assessing exposure.We aimed to evaluate a simple method for assessing outdoor air pollutant concentrations in Oslo, Norway, through a land-use regression method.Samples of nitrogen oxides (NO_x) were collected in two different weeks using Ogawa passive diffusion samplers simultaneously at 80 locations across Oslo. Independent variables used in subsequent regression models as predictors of the pollutants were derived using the Arc 9 geographic information system (GIS) software. Indicators of land use, traffic, population density, and physical geography were tested.The final regression model yielded an adjusted coefficient of determination (R²) of 0.77 for nitrogen dioxide (NO₂), 0.66 for nitric oxide (NO), and 0.73 for NO_x.The results suggest that a good predictive exposure model can be derived from this approach, which can be used to estimate long-term small-area variation in concentrations for individual exposure assessment in epidemiological studies in a highly cost-effective way. These small-area variations in traffic pollution are important since they may have associations with health effects.     

An analytical solution to contaminant diffusion in semi-infinite clayey soils with piecewise linear adsorption

The adsorption of contaminants onto soil particles typically is nonlinear if the contaminant concentration is sufficiently high. A simplified piecewise linear adsorption isotherm consistent with experimental results is proposed as an approximation for nonlinear adsorption behavior. This approximation allows for the use of analytical solution to model solute diffusion of contaminants that exhibit nonlinear adsorption. A moving boundary is introduced to represent significant changes in the retardation factor of clay with an increase in solute concentration. The proposed analytical solutions were validated using experimental data presented in the literature. There is negligible difference between the results obtained by the proposed analytical solution and those obtained by the linear model when C_m/C₀ reached 0.5. The results also show that the model based on linear adsorption using the initial secant of the Freundlich isotherm leads to significantly lower estimated breakthrough time for the contaminant of interest than that obtained using the proposed model. The earlier breakthrough is due to an under-estimation of the amount of adsorption. The proposed method is relatively simple to apply and can be used for evaluating experimental results and verifying more complex numerical models.     

A prediction-based approach to modelling temporal and spatial variability of traffic-related air pollution in Montreal, Canada

Concentrations of traffic-related air pollution can be highly variable at the local scale and can have substantial seasonal variability. This study was designed to provide estimates of intra-urban concentrations of ambient nitrogen dioxide (NO₂) in Montreal, Canada, that would be used subsequently in health studies of chronic diseases and long-term exposures to traffic-related air pollution. We measured concentrations of NO₂ at 133 locations in Montreal with passive diffusion samplers in three seasons during 2005 and 2006. We then used land use regression, a proven statistical prediction method for describing spatial patterns of air pollution, to develop separate estimates of spatial variability across the city by regressing NO₂ against available land-use variables in each of these three periods. We also developed a “pooled” model across these sampling periods to provide an estimate of an annual average. Our modelling strategy was to develop a predictive model that maximized the model R². This strategy is different from other strategies whose goal is to identify causal relationships between predictors and concentrations of NO₂.Observed concentrations of NO₂ ranged from 2.6 ppb to 31.5 ppb, with mean values of 12.6 ppb in December 2005, 14.0 ppb in May 2006, and 8.9 ppb in August 2006. The greatest variability was observed during May. Concentrations of NO₂ were highest downtown and near major highways, and they were lowest in the western part of the city. Our pooled model explained approximately 80% of the variability in concentrations of NO₂. Although there were differences in concentrations of NO₂ between the three sampling periods, we found that the spatial variability did not vary significantly across the three sampling periods and that the pooled model was representative of mean annual spatial patterns.     

Microscale air quality impacts of distributed power generation facilities

The electric system is experiencing rapid growth in the adoption of a mix of distributed renewable and fossil fuel sources, along with increasing amounts of off-grid generation. New operational regimes may have unforeseen consequences for air quality. A three-dimensional microscale chemical transport model (CTM) driven by an urban wind model was used to assess gaseous air pollutant and particulate matter (PM) impacts within ~10 km of fossil-fueled distributed power generation (DG) facilities during the early afternoon of a typical summer day in Houston, TX. Three types of DG scenarios were considered in the presence of motor vehicle emissions and a realistic urban canopy: (1) a 25-MW natural gas turbine operating at steady state in either simple cycle or combined heating and power (CHP) mode; (2) a 25-MW simple cycle gas turbine undergoing a cold startup with either moderate or enhanced formaldehyde emissions; and (3) a data center generating 10 MW of emergency power with either diesel or natural gas-fired backup generators (BUGs) without pollution controls. Simulations of criteria pollutants (NO₂, CO, O₃, PM) and the toxic pollutant, formaldehyde (HCHO), were conducted assuming a 2-hr operational time period. In all cases, NO_x titration dominated ozone production near the source. The turbine scenarios did not result in ambient concentration enhancements significantly exceeding 1 ppbv for gaseous pollutants or over 1 µg/m³ for PM after 2 hr of emission, assuming realistic plume rise. In the case of the datacenter with diesel BUGs, ambient NO₂ concentrations were enhanced by 10–50 ppbv within 2 km downwind of the source, while maximum PM impacts in the immediate vicinity of the datacenter were less than 5 µg/m³.

Implications: Plausible scenarios of distributed fossil generation consistent with the electricity grid’s transformation to a more flexible and modernized system suggest that a substantial amount of deployment would be required to significantly affect air quality on a localized scale. In particular, natural gas turbines typically used in distributed generation may have minor effects. Large banks of diesel backup generators such as those used by data centers, on the other hand, may require pollution controls or conversion to natural gas-fired reciprocal internal combustion engines to decrease nitrogen dioxide pollution.     

Ozone Air Quality over North America: Part II—An Analysis of Trend Detection and Attribution Techniques

ABSTRACT

Assessment of regulatory programs aimed at improving ambient O₃ air quality is of considerable interest to the scientific community and to policymakers. Trend detection, the identification of statistically significant long-term changes, and attribution, linking change to specific clima-tological and anthropogenic forcings, are instrumental to this assessment. Detection and attribution are difficult because changes in pollutant concentrations of interest to policymakers may be much smaller than natural variations due to weather and climate. In addition, there are considerable differences in reported trends seemingly based on similar statistical methods and databases. Differences arise from the variety of techniques used to reduce nontrend variation in time series, including mitigating the effects of meteorology and the variety of metrics used to track changes. In this paper, we review the trend assessment techniques being used in the air pollution field and discuss their strengths and limitations in discerning and attributing changes in O₃ to emission control policies.     

A Simulation Model for Air Pollution over Connecticut

A different approach to mathematically modeling large-scale atmospheric processes is presented. Whereas past approaches have been to develop a model based on an accumulation of information from a specific geographical area, resulting in a model applicable to that area only, we have developed a general mathematical model applicable to any geographical area. The model’s applicability is controlled by specifying the input information describing the meteorological situation and pollution source configuration. A rectangular array of grid points is used to specify both the wind field, by using stream functions, and the average source strength of some pollutant for the area represented by the grid. The diffusion problem is divided into two areas: transport by the mean wind field, and dispersion based on travel time and distance as described by empirical equations. Trajectories of pollutants are traced backwards from the points of interest in the course of the calculations and the contributions of all sources that affect the points of interest are accumulated. The model requires an array of source strength information. An inventory of pollution sources in the State of Connecticut was compiled and maps of source strengths were prepared for five pollutants on a 5000-ft grid-square array. Maps of sulfur dioxide and carbon monoxide source strengths are presented with the resulting concentration distribution for “typical” meteorological conditions. The model permits the changing of meteorological or source values at predetermined intervals so that diurnal changes are incorporated in the calculations. The model has not been verified, but the values of pollution concentration are the right order of magnitude and the resulting patterns are as expected.     

Chemical transport models

Background, aim, and scope  Improving the parameterization of processes in the atmospheric boundary layer (ABL) and surface layer, in air quality and chemical transport models. To do so, an asymmetrical, convective, non-local scheme, with varying upward mixing rates is combined with the non-local, turbulent, kinetic energy scheme for vertical diffusion (COM). For designing it, a function depending on the dimensionless height to the power four in the ABL is suggested, which is empirically derived. Also, we suggested a new method for calculating the in-canopy resistance for dry deposition over a vegetated surface. Materials and methods  The upward mixing rate forming the surface layer is parameterized using the sensible heat flux and the friction and convective velocities. Upward mixing rates varying with height are scaled with an amount of turbulent kinetic energy in layer, while the downward mixing rates are derived from mass conservation. The vertical eddy diffusivity is parameterized using the mean turbulent velocity scale that is obtained by the vertical integration within the ABL. In-canopy resistance is calculated by integration of inverse turbulent transfer coefficient inside the canopy from the effective ground roughness length to the canopy source height and, further, from its the canopy height. Results  This combination of schemes provides a less rapid mass transport out of surface layer into other layers, during convective and non-convective periods, than other local and non-local schemes parameterizing mixing processes in the ABL. The suggested method for calculating the in-canopy resistance for calculating the dry deposition over a vegetated surface differs remarkably from the commonly used one, particularly over forest vegetation. Discussion  In this paper, we studied the performance of a non-local, turbulent, kinetic energy scheme for vertical diffusion combined with a non-local, convective mixing scheme with varying upward mixing in the atmospheric boundary layer (COM) and its impact on the concentration of pollutants calculated with chemical and air-quality models. In addition, this scheme was also compared with a commonly used, local, eddy-diffusivity scheme. Simulated concentrations of NO₂ by the COM scheme and new parameterization of the in-canopy resistance are closer to the observations when compared to those obtained from using the local eddy-diffusivity scheme. Conclusions  Concentrations calculated with the COM scheme and new parameterization of in-canopy resistance, are in general higher and closer to the observations than those obtained by the local, eddy-diffusivity scheme (on the order of 15–22%). Recommendations and perspectives  To examine the performance of the scheme, simulated and measured concentrations of a pollutant (NO₂) were compared for the years 1999 and 2002. The comparison was made for the entire domain used in simulations performed by the chemical European Monitoring and Evaluation Program Unified model (version UNI-ACID, rv2.0) where schemes were incorporated.     

Source apportionment of particulate matter (PM2.5) in an urban area using dispersion,receptor and inverse modelling

Air pollution emission inventories are the basis for air quality assessment and management strategies. The quality of the inventories is of great importance since these data are essential for air pollution impact assessments using dispersion models. In this study, the quality of the emission inventory for fine particulates (PM_2.5) is assessed: first, using the calculated source contributions from a receptor model; second, using source apportionment from a dispersion model; and third, by applying a simple inverse modelling technique which utilises multiple linear regression of the dispersion model source contributions together with the observed PM_2.5 concentrations. For the receptor modelling the chemical composition of PM_2.5 filter samples from a measurement campaign performed between January 2004 and April 2005 are analysed. Positive matrix factorisation is applied as the receptor model to detect and quantify the various source contributions. For the same observational period and site, dispersion model calculations using the Air Quality Management system, AirQUIS, are performed. The results identify significant differences between the dispersion and receptor model source apportionment, particularly for wood burning and traffic induced suspension. For wood burning the receptor model calculations are lower, by a factor of 0.54, but for the traffic induced suspension they are higher, by a factor of 7.1. Inverse modelling, based on regression of the dispersion model source contributions and the PM_2.5 concentrations, indicates similar discrepancies in the emissions inventory. In order to assess if the differences found at the one site are generally applicable throughout Oslo, the individual source category emissions are rescaled according to the receptor modelling results. These adjusted PM_2.5 concentrations are compared with measurements at four independent stations to evaluate the updated inventory. Statistical analysis shows improvement in the estimated concentrations for PM_2.5 at all sites. Similarly, inverse modelling is applied at these independent sites and this confirms the validity of the receptor model results.     

A reanalysis of fine particulate matter air pollution versus life expectancy in the United States

A reduction in population exposure to fine particulate matter air pollution (PM2.5) has been associated with improvements in life expectancy. This article presents a reanalysis of this relationship and comments on the results from a study on the reduction of ambient air PM_2.5 concentrations versus life expectancy in metropolitan areas of the United States. The results of the reanalysis show that the statistical significance of the correlation is lost after removing one of the metropolitan areas from the regression analysis, suggesting that the results may not be suitable for a meaningful and reliable inference.

Implications: The observed loss of statistical significance in the correlation between the reduction of ambient air PM_2.5 concentrations and life expectancy in metropolitan areas of the United States, after removing one of the metropolitan areas from the regression analysis, may raise concern for the policymakers in decisions regarding further reductions in permitted levels of air pollution emissions.     

A Wind Tunnel Study of Gaseous Pollutants in City Street Canyons

Steady state mean concentrations of tracer gas were measured in a 400:1 scale model of an idealized city with variable geometry placed within a wind tunnel at various orientations to the mean flow for a free stream velocity of 6.8 ft/sec. The tracer gas was released from two parallel line sources to simulate lanes of traffic in an effort to quantify the persistence of pollution as well as the mean values realized at street levels. An aerodynamically rough turbulent boundary layer of neutral thermal stratification was employed to simulate the atmosphere. Values of concentration measured in the model city were converted to prototype concentrations in ppm and compared to National Ambient Air Quality Standards. It was shown that single isolated structures may cause favorable mixing of pollution downwind but very high concentrations exist in the immediate leeward vicinity of the building. Two favorable geometries for city blocks tested were found to reduce pedestrian exposure to pollution both near heavy traffic congestion and downwind. It was concluded that the pollutant dilution was controlled by the mean flow rather than by turbulent diffusion and that the lateral spread of the plume was slight as one proceeded downwind of the line source. The combination of favorable geometry and higher dilution velocities may bring pollution levels down to existing Air Quality Standards. The body of information presented in this paper should interest city planners and air quality monitoring personnel, as well as those researchers attempting to study and model flow in city street canyons. It provides order of magnitude estimates on pedestrian and office worker exposure to pollutants under a wide range of conditions.     

Comparison of the performances of land use regression modelling and dispersion modelling in estimating small-scale variations in long-term air pollution concentrations in a Dutch urban area

The performance of a Land Use Regression (LUR) model and a dispersion model (URBIS – URBis Information System) was compared in a Dutch urban area. For the Rijnmond area, i.e. Rotterdam and surroundings, nitrogen dioxide (NO₂) concentrations for 2001 were estimated for nearly 70 000 centroids of a regular grid of 100 × 100 m.A LUR model based upon measurements carried out on 44 sites from the Dutch national monitoring network and upon Geographic Information System (GIS) predictor variables including traffic intensity, industry, population and residential land use was developed. Interpolation of regional background concentration measurements was used to obtain the regional background. The URBIS system was used to estimate NO₂ concentrations using dispersion modelling. URBIS includes the CAR model (Calculation of Air pollution from Road traffic) to calculate concentrations of air pollutants near urban roads and Gaussian plume models to calculate air pollution levels near motorways and industrial sources. Background concentrations were accounted for using 1 × 1 km maps derived from monitoring and model calculations.Moderate agreement was found between the URBIS and LUR in calculating NO₂ concentrations (R = 0.55). The predictions agreed well for the central part of the concentration distribution but differed substantially for the highest and lowest concentrations. The URBIS dispersion model performed better than the LUR model (R = 0.77 versus R = 0.47 respectively) in the comparison between measured and calculated concentrations on 18 validation sites. Differences can be understood because of the use of different regional background concentrations, inclusion of rather coarse land use category industry as a predictor variable in the LUR model and different treatment of conversion of NO to NO₂.Moderate agreement was found between a dispersion model and a land use regression model in calculating annual average NO₂ concentrations in an area with multiple sources. The dispersion model explained concentrations at validation sites better.     

Prediction of daily averaged PM10 concentrations by statistical time-varying model

In this study, a time-varying statistical model, TVAREX, was proposed for daily averaged PM₁₀ concentrations forecasting of coastal cities. It is a Kalman filter based autoregressive model with exogenous inputs depending on selected meteorological properties on the day of prediction. The TVAREX model was evaluated and compared to an ANN model, trained with the Levenberg–Marquardt backpropagation algorithm subjected to the same set of inputs. It was found that the error statistics of the TVAREX model in general were comparable to those of the ANN model, but the TVAREX model was more efficient in capturing the PM₁₀ pollution episodes due to its online nature, therefore having an appealing advantage for implementation.     

Using fugacity to predict volatile emissions from layered materials with a clay/polymer diffusion barrier

Structural insulated panels (SIPs) have significant environmental and energy advantages. However, the tight structure that results may cause degraded indoor air quality and the potential release of volatile organic compounds (VOCs) from these layered materials must be considered. A physically based model for predicting VOC emissions from multi-layer materials is described. Fugacity is used to eliminate the concentration discontinuities at the interface between layers. This avoids an obstacle associated with numerically simulating mass transfer in composite materials. The numerical model is verified for a double-layer system by comparing predicted concentrations to those obtained with a previously published analytical model. In addition, hexanal emissions from multi-layer SIPs are simulated to demonstrate the usefulness of the fugacity approach. Finally, the multi-layer model is used to investigate the impact that clay/polyurethane nanocomposite diffusion barriers can have on VOC emissions. Indoor gas-phase concentrations can be greatly reduced with a barrier layer on the surface, thereby minimizing the environmental impact of SIPs.     

Rural concentrations of nitrogen dioxide pollution throughout Wales

Monitoring of nitrogen dioxide was carried out at over 50 rural sites in Wales throughout 1986. All sites were chosen so as to be remote from any local sources of NO(2) and the values obtained were deemed to be minimum values for the different regions. Measurements were made using a diffusion tube technique which aimed to give mean concentrations of NO(2) for 2-week exposure periods. The results obtained have been used to generate pollution maps to show mean monthly levels of NO(2) for rural environments throughout Wales. It is apparent that levels of NO(2) are generally higher during the winter months. In addition, annual mean concentrations of the pollutant are greatest in the north-eastern and south-eastern parts of Wales with the lowest levels being found along the western coast. The work marks the completion of the first national survey of nitrogen dioxide pollution in Wales. The data are discussed in terms of the potential threat rural concentrations of NO(2) pose to crops and natural vegetation.     

Sulfur dioxide dispersion and source contribution to receptors of downtown Patras, Greece

Background The development of the city of Patras, including harbour relocation, in conjunction with the protection of the regional ecosystems, requires air quality assessment and management. For this reason, a model applicable in the Patras area is necessary and valuable. The goal of this study was to validate a model suitable for predicting the dispersion of sulfur dioxide (SO₂), based on particular activity, topography and weather conditions. Methods We used the US-EPA ISCLT3 integral dispersion model to predict SO₂ concentrations for Patras, Greece. We assumed that the major contribution to Patras air pollution came from central heating, harbour and traffic. We calculated traffic emissions using COPERTIII. Results and Discussion Assigning suitable values of the mixing height, the model predicted the local and spatial distribution of the mean monthly SO₂ concentrations in downtown Patras, as well computed the contribution of the SO₂ emissions originating from each particular source at each receptor location on a seasonal and annual basis. The comparison between predictions and measurements shows that the model performance for estimating the SO₂ concentrations and period pattern is satisfactory. Conclusion The mixing height was the critical parameter for calibrating the model. Model validation promises satisfactory predictions for SO₂ pollution levels on monthly basis. Recommendations and Outlook The model could be used in predicting SO₂ concentrations and source contribution for several downtown Patras receptors using pertinent meteorological and emission information. It could be also extended to predict the dispersion of other primary air pollutants. The calibrated model predictions could be used to fill gaps in monitoring data, saving money and time, and help in assess and manage air quality as Patras develops.     

Surface ozone at rural sites in the latrobe valley and Cape Grim,Australia

Ozone and other air quality data from five rural sites in the industrialized Latrobe Valley, Victoria, have been subject to statistical analyses including linear regression modelling. The behaviour of O₃ in the Latrobe Valley is explained largely in terms of natural background atmospheric processes as observed at Cape Grim, Tasmania.The maximum 1-h average concentration of naturally occurring O₃ (obtained from a 6-year record at Cape Grim) is less than 40 ppb (v/v). In contrast the industrialized Latrobe Valley sites show O₃ values exceeding 40 ppb between 1% and 3% of the time. These higher concentrations occur in conditions consistent with local photochemical production of O₃ via ‘smog’ type processes and appear preferentially at low NO_x concentrations (3–4 ppb) during the afternoon (13–18 h) and at high temperatures (above 25°C).A comparison of observations from an elevated station (750 m) with those from the valley floor shows systematic differences in seasonal and diurnal O₃ variations and the time of day of occurrence of elevated O₃ concentrations which can be explained in terms of the diurnal cycle of convective mixing and mountain/valley winds.A linear regression model incorporating this understanding has accounted for between 43% and 64% of the variance of O₃ concentration at the elevated and rural stations. The statistical model incorporates temperature, time of day, month of year, wind speed, O₃ concentration 24-h earlier and NO_x concentration as variables in the regression equation, with temperature being the dominant variable. The standard deviation of the residual O₃ values (observed minus fitted) is around 5 ppb. Auto and cross correlations are used to show that perhaps half of the unexplained variance is coherent from site to site and hence potentially could be modelled.